Abdul Syafiq Abdull Sukor
Thumbnail Image
Preferred name
Abdul Syafiq Abdull Sukor
Official Name
Abdull Sukor, Abdul Syafiq
Alternative Name
Abdull Sukor, Abdul Syafiq
Sukor, Abdul Syafiq Abdull
Sukor, Abdul Syafiq Bin Abdull
Abdull Sukor, A. S.
Sukor, A. S.A.
Main Affiliation
Scopus Author ID
57209073616
Researcher ID
L-8520-2019

Research Output

Filters

23
Reset filters

Settings
Now showing 1 - 10 of 23
No Thumbnail Available
Publication

Superpixels-based automatic density peaks and fuzzy clustering approach in COVID-19 lung segmentation

2023-12 , Ooi Wei Herng , Aimi Salihah Abdul Nasir , Fatin Nabilah Shaari , Abdul Syafiq Abdull Sukor

Clustering algorithms that rely on minimizing an objective function suffer from the drawback of requiring manual setting of the number of clusters. This limitation becomes particularly evident when applied to image segmentation, where the large number of pixels can lead to memory overflow issues. To overcome this challenge, a reference of Automatic Fuzzy Clustering Framework (AFCF) for image segmentation method has been used as the comparison to the Density Peaks Clustering (DPC) algorithm. AFCF used superpixel algorithm to reduce the spatial information of data during computation, DPC algorithm to generate decision graph, and prior entropy-based fuzzy clustering (PEFC) algorithm to achieve fully automatic segmentation method in determining the number of cluster and the clustering result. In this study, 50 open-source healthy, COVID-19 and pneumonia infected radiographs dataset are acquired from the Kaggle and Github. The radiographs dataset that segmented by DPC is down sampling to 100*100 pixels due to overloading computation. At the end of the image segmentation, a segmentation performance evaluation is conducted based on sensitivity, specificity, accuracy, precision, recall, F-score and time consumed. The result shows that AFCF algorithm has the better overall performance with higher accuracy of 92.48% and F-score 0.9455. Meanwhile, the most highlighted evaluation index is drop to the time consume comparison, AFCF has around 2.7 times faster processing speed compare to DPC algorithm.

View
Thumbnail Image
Publication

Predictive analysis of in-vehicle air quality monitoring system using Deep Learning technique

2022 , Abdul Syafiq Abdull Sukor , Goh Chew Cheik , Latifah Munirah Kamarudin , Xiaoyang Mao , Hiromitsu Nishizaki , Ammar Zakaria , Syed Muhammad Mamduh Syed Zakaria

In-vehicle air quality monitoring systems have been seen as promising paradigms for monitoring drivers’ conditions while they are driving. This is because some in-vehicle cabins contain pollutants that can cause drowsiness and fatigue to drivers. However, designing an efficient system that can predict in-vehicle air quality has challenges, due to the continuous variation in parameters in cabin environments. This paper presents a new approach, using deep learning techniques that can deal with the varying parameters inside the vehicle environment. In this case, two deep learning models, namely Long-short Term Memory (LSTM) and Gated Recurrent Unit (GRU) are applied to classify and predict the air quality using time-series data collected from the built-in sensor hardware. Both are compared with conventional methods of machine learning models, including Support Vector Regression (SVR) and Multi-layer Perceptron (MLP). The results show that GRU has an excellent prediction performance with the highest coefficient of determination value (R2) of 0.97.

View
Thumbnail Image
Publication

Rssi-based for device-free localization using deep learning technique

2020-06-01 , Abdul Syafiq Abdull Sukor , Latifah Munirah Kamarudin , Ammar Zakaria , Norasmadi Abdul Rahim , Sukhairi Sudin , Hiromitsu Nishizaki

Device-free localization (DFL) has become a hot topic in the paradigm of the Internet of Things. Traditional localization methods are focused on locating users with attached wearable devices. This involves privacy concerns and physical discomfort especially to users that need to wear and activate those devices daily. DFL makes use of the received signal strength indicator (RSSI) to characterize the user’s location based on their influence on wireless signals. Existing work utilizes statistical features extracted from wireless signals. However, some features may not perform well in different environments. They need to be manually designed for a specific application. Thus, data processing is an important step towards producing robust input data for the classification process. This paper presents experimental procedures using the deep learning approach to automatically learn discriminative features and classify the user’s location. Extensive experiments performed in an indoor laboratory environment demonstrate that the approach can achieve 84.2% accuracy compared to the other basic machine learning algorithms.

View
Thumbnail Image
Publication

Rssi-based for device-free localization using deep learning technique

2020-06-01 , Abdul Syafiq Abdull Sukor , Latifah Munirah Kamarudin , Ammar Zakaria , Norasmadi Abdul Rahim , Sukhairi Sudin , Nishizaki H.

Device-free localization (DFL) has become a hot topic in the paradigm of the Internet of Things. Traditional localization methods are focused on locating users with attached wearable devices. This involves privacy concerns and physical discomfort especially to users that need to wear and activate those devices daily. DFL makes use of the received signal strength indicator (RSSI) to characterize the user’s location based on their influence on wireless signals. Existing work utilizes statistical features extracted from wireless signals. However, some features may not perform well in different environments. They need to be manually designed for a specific application. Thus, data processing is an important step towards producing robust input data for the classification process. This paper presents experimental procedures using the deep learning approach to automatically learn discriminative features and classify the user’s location. Extensive experiments performed in an indoor laboratory environment demonstrate that the approach can achieve 84.2% accuracy compared to the other basic machine learning algorithms.

View
Thumbnail Image
Publication

Predictive Analysis of In-Vehicle Air Quality Monitoring System Using Deep Learning Technique

2022-10-01 , Abdul Syafiq Abdull Sukor , Cheik Goh Chew , Latifah Munirah Kamarudin , Mao X. , Nishizaki H. , Ammar Zakaria , Syed Muhammad Mamduh Syed Zakaria

In-vehicle air quality monitoring systems have been seen as promising paradigms for monitoring drivers’ conditions while they are driving. This is because some in-vehicle cabins contain pollutants that can cause drowsiness and fatigue to drivers. However, designing an efficient system that can predict in-vehicle air quality has challenges, due to the continuous variation in parameters in cabin environments. This paper presents a new approach, using deep learning techniques that can deal with the varying parameters inside the vehicle environment. In this case, two deep learning models, namely Long-short Term Memory (LSTM) and Gated Recurrent Unit (GRU) are applied to classify and predict the air quality using time-series data collected from the built-in sensor hardware. Both are compared with conventional methods of machine learning models, including Support Vector Regression (SVR) and Multi-layer Perceptron (MLP). The results show that GRU has an excellent prediction performance with the highest coefficient of determination value (R2) of 0.97.

View
Thumbnail Image
Publication

Development of In-situ Sensing System and Classification of Water Quality using Machine Learning Approach

2022-01-01 , Abdul Syafiq Abdull Sukor , Mohamad Naim Muhamad , Ab Wahab M.N.

Quality of water applied to the agriculture sector is one of the factors for agriculture farming to be successful. The use of bad quality irrigation water can cause soil problems. In general, determining water quality model is one of the many interests as it can be used to classify the conditions of water. This project focuses on developing the in-situ sensing system of water quality sensors that can detect parameters of water quality such as pH level, electric conductivity, temperature and total dissolved solid. To validate the approach, there are three types of water samples in a dataset that was collected which include water pipes, soap water and drain water. The types of machine learning models used for classification process are Artificial Neural Network (ANN), Support Vector Machine (SVM), and Decision Tree. The performance showed that SVM model was the highest, ANN was intermediate, and Decision Tree was the lowest. This shows that the SVM model of machine learning approach is the most suitable to be used as the classification model to classify the status of water quality.

View
Thumbnail Image
Publication

Investigation of Different Classifiers for Stress Level Classification using PCA-Based Machine Learning Method

2023-01-01 , Mazlan M.R.B. , Abdul Syafiq Abdull Sukor , Abdul Hamid Adom , Jamaluddin R.B. , Saidatul Ardeenawatie Awang

Undergraduate students experience several changes and face various problems during their time transitioning from adolescence to adulthood. One of the issues during this time is a mental stress disorder. Stress burdens the students either through mental or physical capabilities. The common method of determining stress includes physical examination and clinical diagnosis. However, the method is subjective and time-consuming as doctors need to make sure that their diagnosis is accurate. Thus, the severity of the stress stages could not be easily determined. A new method using machine learning-based algorithms coupled with EEG devices promises to overcome the issues with the current approaches. This paper presents an investigation using machine learning techniques based on Principal Component Analysis (PCA) which allows for the reduction in the dimensionality of datasets to enhance their interpretability while minimizing information loss. The pre-processed EEG data and PCA-based EEG data were compared and analyzed using three machine learning classifiers such as K-Nearest Network (KNN), Naive Bayes (NB) and Multilayer Perceptron (MLP). The results indicate that KNN demonstrated the highest average classification accuracy of 99%, while the other approaches mentioned above averaged around 50% and 80% for NB and MLP respectively. This investigation shows that the KNN classifier is most suitable for the proposed approach.

View
Thumbnail Image
Publication

Pattern Clustering Approach for Activity Recognition in Smart Homes

2022-01-01 , Abdul Syafiq Abdull Sukor , Ammar Zakaria , Latifah Munirah Kamarudin , Wahab M.N.A.

In recent years, studies in activity recognition have shown an increasing amount of attention among other researchers. Activity recognition is usually performed through two steps: activity pattern clustering and classification processes. Clustering allows similar activity patterns to be grouped together while classification provides a decision-making process to infer the right activity. Although many related works have been suggested in these areas, there is some limitation as most of them are focused only on one part of these two processes. This paper presents a work that combines pattern clustering and classification into one single framework. The former uses the Self Organizing Map (SOM) to cluster activity data into groups while the latter utilizes semantic activity modelling to infer the right type of activity. Experimental results show that the combined method provides higher recognition accuracy compared to the traditional method of machine learning. Furthermore, it is more appropriate for a dynamic environment of human living.

View
Thumbnail Image
Publication

Context-aware activity recognition and abnormality detection approaches in smart home environments

2019 , Abdul Syafiq Abdull Sukor

The rising number of elderly population has become a common concern in many countries around the world. The issue has impacted social and economic life of modern societies due to the fact that elderly people are known to suffer from many medical disabilities. As one of the solutions, current technologically-driven approaches, particularly in the area of smart home environments have been developed in recent years to support the independent living and reduce the caregivers’ burden in taking care of elderly individuals. Sensors installed in the environments are used to monitor users’ daily routine to see trends in the behaviour and to be informed of any abnormal activity. However, the accurate interpretation of sensor data in identifying human activities and their abnormal behaviour is still limited. Furthermore, pattern analysis involving these two areas are becoming an increasingly scientific challenge to the real-world environments. This study intends to deal with the issue by investigating appropriate means of pattern recognition and data mining methods within smart home environments. In particular, the study attempts to develop an intelligent reasoning system that can identify residents’ activities and abnormal behaviour of the smart home residents. In this study, two types of activities are identified, i.e., context-related and motion-related activities. The former is classified using the hybrid approach while the latter is performed through the ensemble-based machine learning techniques. The output models produced by these activity recognition approaches are then used as the input for the deep learning networks to produce behavioural model of smart home residents. Experimental procedures are then performed to validate the proposed approach. First, a comparison between the knowledge-driven model and hybrid activity model is carried out to identify the context-related activity. Then, another comparison between the performances of single classifier with multi-classifier system is also performed to identify the motion related activity. Furthermore, for the abnormality detection, several types of reasoning systems are used. These include the case-based reasoning (CBR), deep learning models composed of multi-layer perceptron network (DMLP) and deep recurrent neural network (DRNN) as well as the conventional machine learning algorithms such as naïve Bayes (NB), Support Vector Machine (SVM) and multi-layer perceptron neural network (MLP). The experimental results show that the proposed hybrid approach has better classification rate to identify context-related activity compared to the knowledge-driven model, where the accuracy is obtained at 98.7% ± 0.4. Meanwhile, the multi-classifier system performs better than a single classifier in identifying motion-related activity, with the accuracy of 99.6% ± 0.2. Moreover, DMLP shows higher accuracy rate (98.2%) compared to the DRNN, CBR and other machine learning algorithms for the abnormality detection system. The presented results show that this study can give an impact to the improvement of reasoning process in identifying abnormal situations in smart homes. This can be used in many applications especially in healthcare domains. Furthermore, this study helps to benefit future technologists in order to achieve Society 5.0.

View
Thumbnail Image
Publication

Review of analysis of EEG signals for stress detection

2024-03-07 , Mazlan M.R. , Abdul Syafiq Abdull Sukor , Abdul Hamid Adom , Jamaluddin R.

Mental stress is one of the major contributors to a variety of health issues. Various measures and diagnoses have been created by neurologists and psychiatrists to determine the intensity of mental stress in its early phases. In the literature, several neuroimaging devices and methods for assessing mental stress have been presented. The key candidate chosen is the electroencephalogram (EEG) signal which contains valuable information regarding mental states and conditions. This paper presents reviews of current works on EEG signal analysis for assessing mental stress. The reviews emphasize the significant disparities in the research outcomes and claims of different results from various data analysis methodologies. Accordingly, methods of EEG signals analysis will be used to study the effect of various extracted features and classification methods that associate with mental stress. Apart from that, the utilization of Artificial Intelligence (AI) approach is also investigated to study its significance towards stress detection.

View